Hacker News new | ask | show | jobs
by ajorgensen 1071 days ago
I disagree... follower graph would be pretty important to bootstrap a Twitter like experience. My follower graph on Twitter looks quite different than Instagram or Facebook. The letter also explicitly calls out follower graph and following data which is why I mention it.
3 comments

They used the followers from IG, not Twitter.

Considering who is making this claim and how much they've lied recently, I'd be hesitant to give them the benefit of the doubt.

Simply put, that which can be claimed without evidence can be dismissed without evidence.

I'd estimate less than 25% overlap between my social graph on Twitter and Meta so building a account recommendation system based on my graph from Twitter would be mighty useful. Again this was a just a thought experiment based on two temporally related events and specific wording in a letter nothing more. Of course its on them to prove it.
What use is a follower graph if it doesn't correlate to the accounts on your platform? They have pushed this claim without evidence and Meta has denied them outright. The onus is on Musk's team to produce proof now.
I'm not sure I follow. I'd estimate there's less than a 25% overlap between my Twitter social graph and my Meta (instagram + facebook) social graph. If I was going to build a system to recommend accounts to follow it sure would be handy to know what accounts you followed on the platform most related to the one I'm building. Again I have 0 evidence of anything as I've said and of course its on Musk's team to prove it but just thought the timing and specific wording chosen in the letter was interesting.
I believe

> What use is a follower graph if it doesn't correlate to the accounts on your platform?

Is asking about the point of scraping your Twitter social graph if they can’t actually align it with accounts on the platform. If you follow foo bar on Twitter, it doesn’t mean that foo bar on threads is the same foo bar. The Twitter social graph is useless unless to try to map to meta accounts.

I would venture to guess that 80%+ of the handles are the same on both, enough for the graph to be useful especially for the more popular accounts. If not I bet there’s enough other metadata to do a decent job of mapping them it doesn’t have to be perfect just good enough to get the recommendation engine going. At least enough to bootstrap a recommendation algorithm that is geared more toward primary accounts that create text based content.
Who cares about your guess? You're just doubling down on your previous faulty argument, and anyone would know that there is not a 1:1 between handles on Instagram and Twitter
Meta already has a massive follower graph from using Instagram for signin.

When I logged in to Threads, it added my entire follower list of hundreds from Instagram.

Also, Instagram already links Instagram accounts with Facebook accounts for the majority of users so they graph is even bigger than that (with Facebook friends).

I'm not arguing that... I'm saying that my follower graph for facebook (mostly family) and instagram (mostly friends and diy etc) looks very different than my Twitter graph because the use cases are very different. I would say there is less than 25% overlap between my Twitter social graph and my Meta social graph. If I was bootstrapping a platform like Twitter it would be nice to be able to recommend accounts to follow based on a similar platform (Twitter). Again I have no evidence of anything here... it was literally just two temporally related events that seemed interesting.